Introduction

Child feces are an important source of fecal exposure in household environments and pose a risk to health. Children under five often defecate in or near the household, with research showing young children’s feces are a more common source of fecal contamination than other household members [1, 2]. Evidence also suggests child feces may contain more pathogens than adult feces [3]. Young children themselves are especially at risk of exposure as they actively engage with their surroundings and practice exploratory behaviors such as mouthing [4, 5]. For these reasons, disposal of child feces anywhere other than an improved latrine is considered unsafe [6]. Yet, safe child feces disposal and child latrine use are often not practiced. An analysis of 34 low- and middle-income countries estimated 50.6% of households do not safely dispose of their child’s feces into a latrine or have the child use the latrine [7]. While the world has seen great progress in increasing sanitation access and reducing adult open defecation [8], poor child feces management (CFM) practices remain largely ignored.

In India, safe child feces disposal and child latrine use are especially uncommon practices, despite recent gains in sanitation coverage. Based on the latest National Family Health Survey, only 36% of Indian households with small children disposed of their child’s feces into a latrine or their child used the latrine, even though 61% of households had access to a latrine [9]. The state of Odisha, with 35% latrine coverage, reported the lowest rate with only 13% of households disposing of their child’s feces into a latrine or their child using the latrine [9]. Majorin et al. [11] examined the CFM practices among households in rural Odisha and found that instead of practicing disposal of child feces into a latrine or child latrine use, the majority of children <5 defecated on the ground inside the household or around the household compound (64.8%) and their feces ended up in the household solid waste pile (60.7%). In another cross-sectional study among urban slums in Odisha, most children <5 defecated directly on the ground and their feces were disposed of in the household garbage or in a canal/drain [10]. In both the rural and urban settings, only about one fifth of child feces ended up in the latrine (22% and 25%, respectively) with the majority of this from older children actually defecating in a latrine [10, 11]. There is a clear need to better understand the factors at play for caregiver safe disposal and child latrine use in order to design effective behavior change programming and ultimately improve health.

Much of the CFM literature to date focuses on the broad behavioral term of “safe disposal” defined by the WHO/UNICEF Joint Monitoring Programme [6]. However, this term was derived through a health and measurement lens and in actuality encompasses two distinct “safe” behaviors: child uses the latrine or caregiver puts/rinses the child’s feces into the latrine [6]. From a behavioral lens, it is imperative to disentangle this term and examine the two behaviors separately. While both behaviors are important for reducing fecal exposure, there are likely different factors influencing each that must be uncovered in order to actualize behavior change and subsequent health gains. Here we will use the phrase “caregiver safe disposal” to describe the behavior of the caregiver herself disposing of her child’s feces into a latrine. This behavior consists of a singular actor—the caregiver. In contrast, “child latrine use” refers to the child themselves using the latrine for defecation and it is the child who is the behavioral actor. An area severely underexplored in the CFM literature is the preceding behavior of “child latrine training.” In order for a child to use the latrine, they must first undergo a learning behavioral process whereby the caregiver teaches the child, to at least some degree of intensity, how to use the latrine. This behavior of “child latrine training” is a dyad behavior that involves both the child and the caregiver. In this study we separately assess caregiver safe disposal and child latrine training, with both examinations done from the perspective of the caregiver.

Health behavior change theories outline a variety of factors one must explore in order to diagnose and better understand a behavior. Several theories have been applied to water, sanitation and hygiene (WASH) behaviors such as the integrated behavioral model for WASH (IBM-WASH), behavior centered design (BCD), and RANAS model [12–14]. All highlight the need to assess both contextual and psychosocial factors related to behaviors. Contextual factors include personal characteristics and aspects of the environment which may influence behavior. The environment is often viewed as physical elements that enable or impede behavior but it can also include social elements; for example, the availability of social support. Contextual factors are typically not the focus of behavioral interventions, but an understanding of these factors can better tailor strategies to certain participants and also reveal external barriers to behavioral performance. Psychosocial factors encompass the cognitive elements that influence behavior and have potential for change, such as attitudinal beliefs and normative expectations. The RANAS model in particular outlines five specific psychosocial factors which make-up the theory’s acronym: Risks, Attitudes, Norms, Abilities, and Self-regulation [14].

The past decade has seen a surge in studies that examine contextual factors associated with safe disposal of child feces. Many of these studies analyzed Demographic Health Survey (DHS) data and study settings included India, Bangladesh, Cambodia, and sub-Saharan Africa. Taken together, this body of research has uncovered several common contextual factors related to household, caregiver, and WASH characteristics. Studies show that households are more likely to practice safe child feces disposal when they are in urban settings [15–18], wealthier [15–23], belong to certain religions [10, 16, 17, 23], and when household members practice latrine use [10, 22, 24–26]. At the caregiver level, studies found older mothers [17, 18, 23, 26] who are literate or have higher levels of education [10, 15, 16, 19, 22, 23] and greater exposure to media [16, 17, 21] are more likely to practice safe disposal. For the WASH environment, households with a latrine or water within their household compound; access to an improved latrine or water source in particular; or water was observed at a place for handwashing, were also more likely to safely dispose of their child’s feces [10, 15–17, 19, 21, 23, 25, 27]. However, some contextual factors identified as associated with “safe disposal” may in fact be associated with only one of the two behaviors this broader term covers. For example, many studies found older or ambulatory children are more likely to have their feces safely disposed [10, 11, 15, 17, 18, 21, 24–26], but these characteristics may be specifically associated with child latrine use rather than caregiver disposal of child feces into a latrine.

Compared to contextual factors, there is less research on the psychosocial factors that influence safe CFM practices; still, some shared themes have emerged from qualitative studies. A common finding is a lack of risk perception. Some caregivers do not perceive infant feces to be harmful and as such, do not think they need to be disposed of in a latrine. This low risk perception is attributed to infants’ breastmilk-only diet which makes their feces not smell and appear light in color. However, the feces of older children are often perceived as a health risk because when children transition to solid foods their feces take on the more typical odor and appearance of adult feces [20, 28–31]. Consequently, another common finding is the negative attitudes and social norms caregivers experience when their child defecates around the household environment. Many studies documented the disgust caregivers feel towards seeing their child’s feces and also the social judgement they face, or anticipate, from having a dirty home [28, 31–33].

A few qualitative studies specifically explored caregivers’ experiences with potty training and child latrine use. The studies uncovered several psychosocial factors that acted as barriers to behavioral performance. When it came to potty or toilet training, some caregivers held negative attitudes towards the practice because of its many perceived costs: training is difficult, it takes time, and there are other competing tasks that the caregiver needs to prioritize [26, 34]. The latrine itself is also sometimes perceived as a dangerous environment for small children. Caregivers fear their child might slip or fall into the squat hole and injure themselves [26, 34]. The latrine is also seen as a contaminated place that may make their small children sick [34].

This study aims to quantitatively examine contextual and psychosocial factors influencing caregiver safe disposal and, separately, child latrine training among households in rural Odisha, India. To our knowledge, this is the first quantitative study that assesses behavioral factors for these distinct CFM practices. We first assessed the prevalence of caregiver safe disposal and child latrine use, stratified by child age group, to understand the current state of these behaviors. We then examined personal, physical, and social contextual factors associated with caregiver safe disposal intention and child latrine training. Finally, we applied the RANAS model to more deeply examine psychosocial factors for each behavior. Findings from the study helped inform the design of a behavior change intervention for the Odisha-based NGO Gram Vikas.

Methods

Survey and measures

The survey consisted of six sub-questionnaires: participant and household demographics, caregiver perceived stress, child demographics and child feces management practices, psychosocial factors related to latrine training/disposal, received social support with latrine training/disposal, and characteristics of household water and sanitation facilities. The survey was followed by a structured spot-check of the household’s latrine to assess condition. There were two versions of the psychosocial and social support sub-questionnaires of which caregivers only received one. If the caregiver perceived her child too young to learn to use the latrine, she answered the child feces disposal versions. If the caregiver was currently teaching her child to use the latrine or perceived her child to be old enough, she answered the latrine training versions.

CFM practices.

Caregivers were asked a series of questions about what happened the last time their child defecated in order to measure their CFM practice. If the caregiver had more than one child <5 years old, then the questions were repeated for each child. Additionally, if the caregiver did not know about the last time, which was rare (n = 7; <1%), then they were asked about what happens “usually.” Caregivers were first asked, “The last time the child defecated, where did they defecate?” If the caregiver reported “in the latrine” then this was categorized as child latrine use. If not, a follow-up question was asked: “Where were the child’s feces disposed?” If the caregiver reported “in the latrine” then this was categorized as caregiver safe disposal. All other responses were categorized as unsafe disposal.

Safe disposal intention.

Since caregiver safe disposal was rare (10%, see Table 3), we decided to examine caregiver’s safe disposal intention. Behavioral intention indicates how motivated and willing a person is to perform a behavior, and is a necessary precursor for behavioral action [36]. Safe disposal intention was measured by a self-reported question in the psychosocial factors sub-questionnaire—“How strongly do you intend to always dispose of your child’s feces into the latrine?”—with a 5-point Likert scale response from 1 (not at all) to 5 (very strongly).

Latrine training intensity.

Latrine training intensity, defined as how much effort a caregiver was giving to teaching her child how to use the latrine, was measured based on two self-reported questions. First, caregivers were asked “Are you currently teaching your child how to use the latrine for defecation?” with different clarifying response options (yes, no—my child is not old enough, no—my child already knows how to use the latrine, no—though my child is old enough). This question assessed whether the caregiver herself perceived she had initiated the latrine training process with her child. If the caregiver reported “yes,” then a follow-up was asked—“During the last week, when your child needed to defecate, how often did you take your child to the latrine and teach them how to use it?”—with a 5-point Likert scale response from 1 ((almost) never (0%)) to 5 ((almost) always (100%)). Latrine training intensity was thus measured on a 0 to 5 scale based on the response to the second question and with 0 assigned to caregivers who reported “no-though my child is old enough” to the first question. Caregivers who perceived their child was not yet old enough to learn how to use the latrine or who reported their child already knew how to use the latrine were not assigned a latrine training intensity value. If the caregiver had multiple children <5 years old who she was teaching how to use the latrine or she perceived old enough to learn, then responses for the youngest child were used in the analysis.

Contextual factors.

The contextual factors examined included personal, social, and physical characteristics: personal characteristics of the caregiver, child and household; characteristics of the amount and type of social support that the caregiver received; and physical characteristics of the household’s water and latrine infrastructure.

Caregiver characteristics were self-reported and included age, number of years of education, whether or not the caregiver was employed/self-employed, and whether or not the caregiver used the latrine the last time she defecated.

Child characteristics included the child’s age, sex, and developmental abilities. For developmental abilities, the caregiver was asked to report “yes” (1) or “no” (0) if her child was doing any of the following: walk on their own, squat on their own, speak in full sentences, follow directions, and eat adult foods. The developmental abilities were analyzed individually as dichotomous variables. However, in the latrine training intensity analysis, since almost all children were able to walk (97%) and ate adult foods (98%) these were excluded from the analysis. In the safe disposal intention analysis, only the ability to walk was examined. For caregivers with multiple children <5 years old, the characteristics of the youngest child were included in the analysis.

Household characteristics included religion, caste group, whether or not the household had multiple children <5 years old, the number of household members that helped with childcare, and wealth. Household wealth was measured by constructing an asset index (scooter/motorcycle, television, telephone [landline or mobile], refrigerator, mattress, cot, chair, table, sewing machine, pressure cooker, watch/clock, electric fan, water pump, animal drawn cart, thresher, tractor, electricity, livestock) and using polychloric principles component analysis to categorize households into five wealth quintiles (quintile 1 –least wealthy; quintile 5 –most wealthy) [37].

The amount of social support the caregiver received with latrine training/disposal was measured using the received social support sub-questionnaire. The questionnaire examined three types of support—emotional, instrumental, and informational. The items were adapted from several validated metrics [38–42] with several items examining one type of social support. In each item, the caregiver was asked about a specific supportive act she experienced in the last week with a 6-point Likert scale response from 1 (completely disagree) to 6 (completely agree). Many of the supportive acts used in the items came from real examples provided by caregivers in prior qualitative research. A score was constructed for each of the three types of social support by taking the average of their relevant items. A higher score indicates greater received support. Example items and the Cronbach’s alpha internal reliability coefficient for each constructed score are presented in Tables 1 and 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Description of psychosocial and social support factors included in the models for caregiver safe disposal.

https://doi.org/10.1371/journal.pone.0274069.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Description of psychosocial and social support factors included in the models for child latrine training.

https://doi.org/10.1371/journal.pone.0274069.t002

Household water and latrine characteristics were either reported by the caregiver or observed by the surveyor. Piped water access was measured as the number of hours in the last 24 hours that caregivers reported they could not get water from their private tap. If the caregiver’s household did not have piped water (n = 56; 7%), then a value of 24 was assigned. For the household latrine, caregivers reported on the number of pits and whether or not the latrine was currently used for defecation. However, the latter was dropped from the analysis since almost all reported the latrine to be in use (93%). Surveyors observed the location of the latrine (“in/within 50 feet of the household compound” (1) or “>50 feet” (0)) and if it had functional piped water (“tap in latrine and water comes out when tested” (1) or “no tap/water does not come out” (0)). Lastly, we determined if the latrine structure was "fully intact” (1) or “not fully intact” (0) based on surveyor observations. The surveyor recorded “yes” or “no” if the roof, walls, floor, and door of the latrine were intact, if the walls were at least 5 feet high, and if the squat pan could be used. If “yes” was answered for all items then the latrine structure was considered “fully intact.” Pit condition was not included because most surveyors recorded pits as “not visible” since they were often below ground.

Psychosocial factors.

The psychosocial sub-questionnaires were based on the RANAS model and heavily informed by qualitative research. Items discussed both the desired behavior (i.e. disposal of child feces into the latrine or teaching child how to use the latrine) and the undesired behavior (i.e. disposal of child feces outside or child defecating outside). Most items used a 5-point Likert response scale but a few items were dichotomous. The safe disposal questionnaire included a “yes” (1) or “no” (0) item that examined village injunctive norms and an item on barrier planning was turned dichotomous with 0 indicating no plan and 1 indicating a plan. The latrine training questionnaire included one item that examined caregivers’ personal beliefs about the age at which mothers should start teaching their child how to use a latrine with different ages as response options. This item was turned dichotomous where 1 indicated the caregiver believed children < = 2 years old should be taught and 0 indicated >2 years old. Some 5-point Likert items were combined into a single constructed factor by averaging the responses. All factors were coded so that a higher value indicated greater favorability of the desired behavior. Example items and the Cronbach’s alpha internal reliability coefficient for constructed factors are presented in Tables 1 and 2.

Results

A total of 1033 caregivers met eligibility and were asked to participate in the survey. Among these, 41 (4%) did not consent, 113 (11%) were secondary caregivers, 10 (1%) were male primary caregivers, 42 (4%) had a child that had already completed their latrine training, and 36 (3%) ended the survey early. This resulted in 791 primary female caregivers included in the analysis.

Child defecation and feces disposal practices

The caregivers reported on the CFM practice for 906 children <5 years old (Table 3). There was a roughly equal distribution of children across the age groups from 0 to <60 months and 48% of children were female.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Child feces management practices the last time the child defecated, stratified by child age.

https://doi.org/10.1371/journal.pone.0274069.t003

At their last defecation event, 38% of children used the latrine, an additional 10% had their feces safely disposed of into the latrine, and the remaining 52% had their feces unsafely disposed. Among children who had their feces unsafely disposed, the most common practices were to dispose of the feces into an open area away from the household compound (29%), into the backyard of the household (19%), or into the household garbage pile (18%). Among children who used the latrine, almost all (93%) were two years old or older, and this age group was also the least likely to have their feces safely disposed of when defecating outside the latrine (3.1%). Accordingly, among children who did have their feces safely disposed of, albeit this was rare, most (81%) were less than two years old.

We examined the defecation and disposal practices by analysis group: latrine training intensity and safe disposal intention. As expected, for the children of caregivers in the latrine training intensity analysis the majority (57%) used the latrine the last time they defecated. For children of caregivers in the safe disposal intention analysis, only 15% had their feces safely disposed the last time; again, highlighting this is an uncommon behavior in the study population. Children in the latrine training intensity group were on average about three years old (M = 38.87 months, SD = 11.49) while children in the safe disposal intention group were on average about one year old (M = 15.09 months, SD = 10.00). Almost all children in the latrine training intensity group could walk (97%) compared to about half in the safe disposal intention group (54%). Children in the latrine training intensity group could also squat on their own (87%), follow directions (87%), and speak in full sentences (79%).

Contextual factors

Caregiver safe disposal intention.

The contextual factors multivariate linear regression model for safe disposal intention identified household wealth (b = 0.109; 95% CI 0.003, 0.215), latrine location (b = -0.645; 95% CI -1.042, -0.247), and informational support (b = 0.098; 95% CI 0.002, 0.194) as significant predictors (Table 4). Caregivers from wealthier households and who received more informational support on CFM had a greater intention to practice safe disposal of their child’s feces. Caregivers who had a latrine in or near the household compound (within 50 feet) had weaker safe disposal intention compared to caregivers who had a latrine >50 feet from the household compound. The model explained only 8% of the variance.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Multivariate linear regression models for caregiver safe disposal intention.

https://doi.org/10.1371/journal.pone.0274069.t004

Child latrine training.

The contextual factors multilevel mixed effects regression model for latrine training intensity identified caregiver latrine use (b = 1.307; 95% CI 0.877, 1.737), latrine structure fully intact (b = 0.452; 95% CI 0.021, 0.883), and instrumental support (b = 0.212; 95% CI 0.075, 0.349) as significant predictors (Table 5). Caregivers who used the latrine the last time they defecated, had a fully intact latrine with usable squat plate, and who received greater instrumental support, were more intensely teaching their child how to use the latrine. While the model explained a modest level of the overall and within village variance (19% and 14%, respectively), it explained a high level of the variance between caregivers living in different villages (48%).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Multilevel mixed effects linear regression models for latrine training intensity.

https://doi.org/10.1371/journal.pone.0274069.t005

Discussion

We aimed to descriptively examine child latrine use and feces disposal practices, and subsequently identify contextual and psychosocial factors associated with two specific behaviors: caregiver safe disposal and child latrine training. Since caregiver safe disposal was found to be a rare practice—only 10% among children <5—we instead examined caregiver safe disposal intention. With regard to contextual factors, we found caregivers had stronger behavioral intention to safely dispose when they came from wealthier households and had greater informational support, but less intention when the latrine was in/near the household compound. In contrast, caregivers more intensely practiced latrine training with their child when they themselves used the latrine for defecation, had a fully intact latrine, and greater instrumental support. With regard to psychosocial factors, caregivers had stronger intention to safely dispose when their households expected them to do so, they felt strongly committed to the behavior, and had a plan for what to do when faced with a water shortage barrier. For latrine training, caregivers more intensely taught their child how to use the latrine when they believed their child was at risk of becoming sick if they practiced OD; viewed child OD as unbeneficial; liked teaching their child; personally felt it was important for the child’s father to help; felt confident in their ability to teach their child; and had greater action control over their training practice. Interestingly, caregivers put less effort into latrine training when they felt more concerned for their child’s safety when the child defecated outside. These findings underscore the need to separately assess unique CFM practices in order to uncover the different behavioral factors at play. Moreover, our results offer a road map for practitioners working in similar settings on the types of behavior change strategies to consider in their CFM programming.

While many studies also showed household wealth and caregiver latrine use as associated with the broader term of “safe disposal” [10, 15–26], we parsed out these factors to reveal they are behaviorally specific: household wealth with caregiver safe disposal intention and caregiver latrine use with child latrine training. Household wealth may act as a proxy for a better quality latrine or exposure to more urban settings, such as a household member who works in a city, where caregiver safe disposal is more commonly practiced [9]. As for latrine training, it is intuitive that caregivers who use a latrine themselves would be more likely to then teach their child to use the latrine; this can also be viewed as a strategy in childrearing to model desired behaviors [43]. Interestingly, we also found caregivers have stronger intentions to safely dispose of their child’s feces when their latrine is farther away from the household. Caregivers might prefer this greater distance at a specific child development stage: when their child is not yet ready to use the latrine but has started to eat adult foods and their feces now smell. However, since the vast majority of caregivers had a latrine in/near their household, it should be noted that only a few caregivers may have been driving this result. These findings emphasize the need to separately examine distinct CFM behaviors to reveal a clearer picture. Furthermore, for programming, it might prove harder to promote behavioral adoption of safe CFM practices in areas where households are poorer and there is less latrine use among adults.

We also uncovered different results by behavior for social support. Social support is an important resource in a person’s social environment; it helps strengthen one’s self-confidence and acts as a “buffer” against stressors [44, 45]. These differential findings are consistent with the stress-support matching hypothesis: a stressor must be matched with the type of support that is specifically needed [46]. The behaviors of safe disposal and child latrine training may be viewed in this light, as stressors in caregivers’ daily lives among the myriad of other childcare and household responsibilities. Our results suggest caregivers desire informational support—advice and guidance—when it comes to safely disposing of their child’s feces but desire instrumental support—acts of assistance—when it comes to latrine training. The desire for instrumental support aligns with the personal norm and self-efficacy psychosocial factors also found to be significantly associated with latrine training. Caregivers put greater effort into latrine training when it was personally important to them that the child’s father help, potentially indicating that they were receiving this instrumental support from their husband. Caregivers also more intensively taught their child when they were more confident; in a separate examination using mediation analysis we show how instrumental support actually aids latrine training through bolstering a caregiver’s perceived self-efficacy [47]. As the WASH field recognizes how overburdened women are in taking up improved WASH practices, programmatic strategies that bolster support in the household or among other social networks could prove key.

One psychosocial factor associated with caregiver safe disposal intention related to injunctive norms. Descriptive norms are beliefs that others perform a certain behavior while injunctive norms are beliefs that others expect you to also perform this behavior [48]. When there is both a descriptive and injunctive norm in place, the behavior is deemed to be a “social norm” [48]. In this study we found caregivers held stronger safe disposal intentions when others in their household expected them to do this practice. Interestingly, we did not find a significant association with the village descriptive norm factor. Other qualitative studies have documented the role of both descriptive and injunctive norms around safe disposal. In urban Bangladesh, caregivers of children <3 years old explained that the common practice in their neighborhood (i.e. descriptive norm) was to throw child feces and diapers into open areas [33]. In western Kenya, caregivers of children <2 years old similarly emphasized it was a social norm to bury feces, dispose of wash water (from cleaning cloths/nappies) wherever was convenient, and there was no expectation for older siblings, who sometimes cared for their younger sibling, to practice safe disposal [28]. Taken together, these findings suggest interventions that aim to improve caregiver safe disposal should consider behavior change techniques that influence norms, such as public commitments or indicators of others’ approval [49].

The other significant psychosocial factors for caregiver safe disposal intention were barrier planning and commitment, which are self-regulation factors. Self-regulation is about a person’s ability to continue performing a behavior in spite of conflicting priorities, distractions, and other barriers [14]. As qualitative studies document, caregivers are busy with many tasks and find it difficult to incorporate new CFM practices [31, 33]. As such, along with norm techniques, CFM interventions focused on safe disposal can incorporate self-regulation techniques like goal setting and prompting caregivers to cope with barriers [49]. These results again suggest that strengthening informational support may be useful too: programs could engage a community mobilizer or family member to provide supportive guidance to caregivers on how to navigate different barriers with practicing safe disposal, such as water shortages, time and workload.

In contrast to caregiver safe disposal, child latrine training was associated with a wider variety of psychosocial factors. Along with norms and self-regulation, child latrine training was also associated with attitudes, self-efficacy, and risk perceptions; all of which require different types of behavior change techniques. For attitudes, caregivers more intensively taught their child how to use the latrine if they liked latrine training and viewed child open defecation as unbeneficial. Behavioral programs could include techniques that demonstrate latrine training as an enjoyable process, such as highlighting the positive emotions a caregiver might feel from training successes, and other techniques that illustrate the costs of the undesired behavior (i.e. child practicing open defecation). When it comes to improving caregivers’ self-efficacy, some behavior change techniques include organizing social support, particularly instrumental support as already discussed, modeling or demonstrating the behavior, and provision of hardware [49]. There are mixed reviews on hardware like potties—some caregivers find them helpful while others do not [30, 33]. Novel hardware for latrine training may be needed, especially in contexts where children must learn to squat rather than sit when defecating. Lastly, caregivers more intensely taught their child how to use the latrine if they more strongly believed their child was at risk of becoming sick if they defecated outside. Behavioral programs could apply techniques that inform caregivers of the severe health risks that open defecation poses to their child.

Surprisingly, caregivers put less effort into latrine training when they felt more concerned for their child’s safety when the child defecated outside. It could be that this risk perception serves as a proxy for when a caregiver perceives her child old enough to use the latrine. A few studies documented that caregivers perceive the latrine to be a dangerous place for small children, risking physical injury and illness [26, 34]. If a caregiver feels more concerned for her child’s safety when they go for open defecation, she may also feel concerned about them using the latrine. However, this finding may have resulted from a type 1 error since we tested a large number of factors and the p-value was much higher compared to the other significant factors.

This study has some limitations. Firstly, as part of the eligibility criteria for the trial, all caregivers engaged had a household latrine and all villages had previously participated in the Gram Vikas MANTRA community-based water and sanitation intervention. As such, we caution against generalizing these findings to contexts where households do not have a strong enabling environment (i.e. access to a latrine and water) for practicing safe disposal and latrine training. Secondly, due to its uncommon practice, we were unable to directly examine the behavior of caregiver safe disposal and instead had to assess the behavioral precursor of intention. It is possible that different contextual and psychosocial factors are related to the behavior itself compared to intention; however, the two are viewed as closely linked and this approach has been used before [50]. Lastly, psychosocial factors related to child latrine training were examined solely from the perspective of the caregiver but different factors may be important from the child’s perspective.

Conclusions

As the water and sanitation field considers a more “transformative WASH” approach to reduce fecal exposure and achieve health gains [51], there is a need to better understand caregiver safe disposal and child latrine training. Our results exemplify it is critical to examine these behaviors separately so that their unique influencing factors are uncovered and more targeted behavioral strategies can be selected. However, while we urge the field to recognize these behaviors as distinct and to avoid the muddling term of “safe disposal,” programmatically these behaviors should be addressed together as they are linked by child development. Caregivers first need to practice safe disposal but then as their child becomes developmentally ready they should transition to latrine training with their child. Our results showed it is around two years old when children start to use the latrine, which aligns with toilet training guidance from the Indian Academy of Pediatrics and others [52–54]. Practitioners can use the findings from this study to develop more holistic CFM programs that address both practices but apply specific behavior change techniques for each; tailor which content a caregiver receives based on her child’s age; and also support those caregivers of younger children navigate the transition to latrine training as their child grows.

Supporting information

Acknowledgments

We would like to thank all of the participants who offered their time in taking the survey, especially when they were busy with caregiving tasks. We would also like to acknowledge Munmun DasMohapatra for translating the survey into Odia and her skillful training of the survey team, Indrajit Samal for providing invaluable logistical support throughout data collection, and Miriam Harter for her expertise and guidance on the piloting process of the psychosocial questionnaires. We are also grateful to the survey team who travelled far to reach all the study villages and had many long days of data collection.